You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be good if we removed the void as it would be an improper binding.
This task should be simple, since we generate out a typing map, in the past we had
;; Update logic so that we can get inference on this.
(-> remove-void-bindings (ir:fully-expanded-list) ir:fully-expanded-list)
(defunremove-void-bindings (terms)
which would scan the syntax manually for applications to remove the bindings.
Thus this should be updated to use the typing map to check if a value is void and do the transformations then.
The text was updated successfully, but these errors were encountered:
@@ -216,7 +216,8 @@ of the user program is preserved." (list (ir:var term)))) (ir:make-multiple-bind :var nil :val (ir:value term))) ((and (typep value 'ir:reference)- (sycamore:tree-set-find set (ir:name value)))+ (or (alu.spec.term-op:void-reference? value)+ (sycamore:tree-set-find set (ir:name value)))) (sycamore:tree-set-insertf set (ir:var term)) nil)
in remove-void-bindings the void is gone, however the shift to using the typing map should happen here, as more primitives should trigger this issue
With the last commit, explciit voids can be generated from standalone
with-constraint
.This means that the code currently generates out a void with the following code
It would be good if we removed the void as it would be an improper binding.
This task should be simple, since we generate out a typing map, in the past we had
which would scan the syntax manually for applications to remove the bindings.
Thus this should be updated to use the typing map to check if a value is void and do the transformations then.
The text was updated successfully, but these errors were encountered: